Publication detail

Using Machine Learning Techniques in the Visual Detection of Starlings in Vineyards

NOVOTNÝ, J.

Original Title

Using Machine Learning Techniques in the Visual Detection of Starlings in Vineyards

English Title

Using Machine Learning Techniques in the Visual Detection of Starlings in Vineyards

Type

conference paper

Language

en

Original Abstract

This project deals with visual detection of starlings. The aim is to create an early warning de-tection system that protects crops from flocks of starlings. This system uses a computer vi-sion and machine learning algorithms. In the first phase, the activity in the vineyard was col-lected. Further, the neural network model using cloud-based AutoML platform was trained. The final system classifies objects into several categories. These categories include individual birds, flocks and various unintended objects such as flies and bees. Overall, the flock detec-tion algorithm achieved 89% accuracy and 94% sensitivity.

English abstract

This project deals with visual detection of starlings. The aim is to create an early warning de-tection system that protects crops from flocks of starlings. This system uses a computer vi-sion and machine learning algorithms. In the first phase, the activity in the vineyard was col-lected. Further, the neural network model using cloud-based AutoML platform was trained. The final system classifies objects into several categories. These categories include individual birds, flocks and various unintended objects such as flies and bees. Overall, the flock detec-tion algorithm achieved 89% accuracy and 94% sensitivity.

Keywords

Starlings detection, Flocks detection, Computer vision, Machine Learning

Released

27.04.2021

Publisher

Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií

Location

Brno, Czech Republic

ISBN

978-80-214-5943-4

Book

EEEICT_2021

Edition

1st

Pages from

102

Pages to

105

Pages count

4

URL

https://www.fekt.vut.cz/conf/EEICT/archiv/sborniky/EEICT_2021_sbornik_2.pdf

Documents

BibTex 


@inproceedings{BUT171587,
  author="Josef {Novotný}",
  title="Using Machine Learning Techniques in the Visual Detection of Starlings in Vineyards",
  annote="This project deals with visual detection of starlings. The aim is to create an early warning de-tection system that protects crops from flocks of starlings. This system uses a computer vi-sion and machine learning algorithms. In the first phase, the activity in the vineyard was col-lected. Further, the neural network model using cloud-based AutoML platform was trained. The final system classifies objects into several categories. These categories include individual birds, flocks and various unintended objects such as flies and bees. Overall, the flock detec-tion algorithm achieved 89% accuracy and 94% sensitivity.",
  address="Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií",
  booktitle="EEEICT_2021",
  chapter="171587",
  edition="1st",
  howpublished="online",
  institution="Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií",
  year="2021",
  month="april",
  pages="102--105",
  publisher="Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií",
  type="conference paper"
}